Kinship Term Generation - Wedding Invitations: leaderboard

Metric: Accepted-term accuracy (%; short wedding invitation to the specified relative, 231 responses per model; fully specified relationship (L3) prompts for 25 kinship relations in Hindi, Tamil and Korean, three runs per prompt at temperature 0.7; a response counts as correct when an accepted native-script kinship term (with listed spelling and honorific variants) appears in it). Source: arxiv.org. Saturation forecast: Around December 2026. 5 models tracked.

Top models

#ModelScore
1GLM-5.171
2Kimi K2.650.22
3GPT-OSS-120B35.06
4Llama 3.3 70B Instruct25.11

Interactive version: theaggregate.ai/benchmark?slug=kinship-term-generation-wedding-invitations · How It Works · Data refreshed daily, snapshot 2026-09-26.